THE SHARK AI PROJECT: UTILIZING PALEONTOLOGY TO CATALYZE LEARNING ABOUT ARTIFICIAL INTELLIGENCE

Artificial Intelligence (AI) is becoming ubiquitous in modern society and advancing rapidly. Consequently, educators are struggling to keep pace with the changing landscape. Our three-year project uses fossil sharks to activate teacher self-efficacy and student interest in learning about AI, specifically machine learning (ML) and computer vision. Three cohorts of Florida middle school teachers, primarily from Title 1 schools, participated in a five-day professional development (PD) workshop hosted at the University of Florida and Florida Museum of Natural History. During this PD, teachers were introduced to a STEM-integrated curriculum, consisting of five scaffolded modules that are aligned with Florida and national K-12 education standards. The curriculum guides students through the process of developing and evaluating computer vision models to identify fossil shark teeth using a free, open-access platform, Google’s Teachable Machine. This process entails classifying fossil shark and ray teeth based on Linnaean taxonomy and functional morphology, developing training datasets reflecting these classification schemes, and then iteratively evaluating and improving their models by assessing the training data for potential sources of bias. The curriculum culminates with the opportunity for students to develop their own unique model, providing a summative assessment of students’ comprehension of the machine learning development process. Over the course of the school year, teachers are supported with quarterly check-ins and scientist role model visits facilitated by the Scientist in Every Florida School program. Through our project evaluation, teachers have indicated increased self-efficacy in terms of teaching AI and paleontology concepts, as well as increased student engagement and STEM career interest.